Question Overheating 5900X with a Lian Li Galahad 360

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
For starters my rig is composed of:

Mobo Asus X570 ROG Crosshair VIII Hero (WI-FI)
CPU AMD Ryzen 9 5900X
RAM Crucial Ballistix Black 2x8GB
GPU Inno3D RTX 3080 iChill x4
PSU FSP Hydro G Pro 750W 80+ Gold
Cooling Lian Li Galahad 360
Case Phanteks P500A DRGB

You could easily notice that I have picked parts to get the best temperatures possible.

I was pretty relieved after initial installation of Lian Li Galahad 360. With the Phanteks case combined, these two alleviated the heat stress.

However, for the last month or two month, on seldom cases, I came across with system shutting down to see CPU Overheat message as I re-run it.

Please check the HWiNFO64 screenshot I grabbed few minutes before, after hanging around in the MENU of Call of Duty Warzone - I did not actually get in-game, mind you. I am getting stutters in-game over time, most likely from CPU getting throttled.



I got all the right parts here, I guess, to be able to use this rig without having to care about any kind of overheating. However, once again, I am in the same hellpit. My overheating problem is like a saga. It just comes back.

What do you have in mind? What could you suggest please? What I did so far were:

  • Resetting PC
  • Formatting PC
  • Re-applying thermal paste
  • Downgrading BIOS to a version I remember were from "good old times"

I am so desperate guys, please help.
 
For starters my rig is composed of:

Mobo Asus X570 ROG Crosshair VIII Hero (WI-FI)
CPU AMD Ryzen 9 5900X
RAM Crucial Ballistix Black 2x8GB
GPU Inno3D RTX 3080 iChill x4
PSU FSP Hydro G Pro 750W 80+ Gold
Cooling Lian Li Galahad 360
Case Phanteks P500A DRGB

You could easily notice that I have picked parts to get the best temperatures possible.

I was pretty relieved after initial installation of Lian Li Galahad 360. With the Phanteks case combined, these two alleviated the heat stress.

However, for the last month or two month, on seldom cases, I came across with system shutting down to see CPU Overheat message as I re-run it.

Please check the HWiNFO64 screenshot I grabbed few minutes before, after hanging around in the MENU of Call of Duty Warzone - I did not actually get in-game, mind you. I am getting stutters in-game over time, most likely from CPU getting throttled.



I got all the right parts here, I guess, to be able to use this rig without having to care about any kind of overheating. However, once again, I am in the same hellpit. My overheating problem is like a saga. It just comes back.

What do you have in mind? What could you suggest please? What I did so far were:

  • Resetting PC
  • Formatting PC
  • Re-applying thermal paste
  • Downgrading BIOS to a version I remember were from "good old times"
I am so desperate guys, please help.
It's not unexpected for temps to max out around 90C when boosting with Ryzen CPU's. The more important thing to look at is the Average temperature. Especially the CPU Die (Average), average as that's an average of the averages: it averages all the sensors together across the die and then averages those across time. Watch that sensor in particular while running a moderately heavy all-core workload, like Cinebench 20 or 23. Reset the average counter by clicking on the clock icon at bottom of sensor screen

Even so, getting an overheat message might indicate a problem. Are you running the pump in the AIO at full speed, constantly? And have you double checked that all fasteners are well tightened on the pump/block?
 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
It's not unexpected for temps to max out around 90C when boosting with Ryzen CPU's. The more important thing to look at is the Average temperature. Especially the CPU Die (Average), average as that's an average of the averages: it averages all the sensors together across the die and then averages those across time. Watch that sensor in particular while running a moderately heavy all-core workload, like Cinebench 20 or 23. Reset the average counter by clicking on the clock icon at bottom of sensor screen

Well that average of averages say it is around 57.9.

About Cinebench 20, ClockTuner for Ryzen (CTR) 2.1 by 1usmus runs it as the starting point of its diagnosis for tuning. A few seconds after the render pass starts, Cinebench quits and returns to CTR which says the test was abandoned since 91 celsius was reached. I remember going through render runs before, not abruptly ending like this.

Even so, getting an overheat message might indicate a problem. Are you running the pump in the AIO at full speed, constantly? And have you double checked that all fasteners are well tightened on the pump/block?

If that means water pump and AIO pump duty cycles being fixed at 100% in BIOS, yes I am. I removed the pump block today to check the thermal paste - removed residue which was still mushy and quite spreaded and re-applied what little Kryonaut I had left and fastened the block after a few tries. Checked the screws being fully turned.

watch your CPU temperatures and voltage live while in a game and during a stress test with RTSS and/or HWinfo OSD.
see what your real-time temps are as you're utilizing the system with high level processing.

Well seeing it hit 90 at Call of Duty Warzone's main menu and remembering how I got stutters yesterday most likely occuring because of throttles, I am pretty sure they hit well above 90 during that.

Btw, I installed Windows 11 as I reformatted my PC and I can't get on-board audio working at all. The whole thing collapsed like a concrete maze on top of me. I got errors while creating W10 USB twice during the last hour - trying it for a third time to revert back to Windows 10.
 
Well that average of averages say it is around 57.9.

About Cinebench 20, ClockTuner for Ryzen (CTR) 2.1 by 1usmus runs it as the starting point of its diagnosis for tuning. A few seconds after the render pass starts, Cinebench quits and returns to CTR which says the test was abandoned since 91 celsius was reached. I remember going through render runs before, not abruptly ending like this.
....
The average is quite good but then it's also reflecting time when the CPU's doing nothing, that's why you reset the average counters (the clock) during a test to see what it is.

It's been a while since I used CTR didn't find it useful for my 3700X. IIRC, it's a tool for finding an overclock and voltage by experimentation. It's probably just telling you the combination it used is overheating which is something very easy to do with a 5900X CPU.

That's just one reason most people don't overclock these CPU's, at least not conventionally with fixed clocks and voltage even when low: they overheat easily if not allowed to lower clocks as needed. Instead use PBO while undervolting with Curve Optimizer which helps keep temperatures much lower because it doesn't ignore the boost algorithm. Actual performance usually ends up very similar as a best-case overclock, and usually better for lightly threaded work loads such as gaming.
 
Last edited:
If that means water pump and AIO pump duty cycles being fixed at 100% in BIOS, yes I am
of course this would only relate if you are actually using one of those two AIO_PUMP or W_PUMP headers.

there's a good chance that since this heat issue has just crept up after running fine for some time that your pump, tubes, and/or radiator are clogged with bubbles or debris.
an easy way to dislodge them is to remove the entire setup and rotate and flex it all about and try and dislodge this "clog".

i haven't noticed you mention how this AIO is mounted.
 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
The average is quite good but then it's also reflecting time when the CPU's doing nothing, that's why you reset the average counters (the clock) during a test to see what it is.

It's been a while since I used CTR didn't find it useful for my 3700X. IIRC, it's a tool for finding an overclock and voltage by experimentation. It's probably just telling you the combination it used is overheating which is something very easy to do with a 5900X CPU.

That's just one reason most people don't overclock these CPU's, at least not conventionally with fixed clocks and voltage even when low: they overheat easily if not allowed to lower clocks as needed. Instead use PBO while undervolting with Curve Optimizer which helps keep temperatures much lower because it doesn't ignore the boost algorithm. Actual performance usually ends up very similar as a best-case overclock, and usually better for lightly threaded work loads such as gaming.

Here is what I got as averages after 5-10 mins of Warzone having reset the counters:



of course this would only relate if you are actually using one of those two AIO_PUMP or W_PUMP headers.

there's a good chance that since this heat issue has just crept up after running fine for some time that your pump, tubes, and/or radiator are clogged with bubbles or debris.
an easy way to dislodge them is to remove the entire setup and rotate and flex it all about and try and dislodge this "clog".

i haven't noticed you mention how this AIO is mounted.

Please see the illustration below for the orientation of fans that are represented by an arrow each.

 
Here is what I got as averages after 5-10 mins of Warzone having reset the counters:





Please see the illustration below for the orientation of fans that are represented by an arrow each.

That's a good mounting arrangement. Any air should settle in up at the top tank and never be able to make it back into the pump.

Yeah...those average temps seem a bit high to me for simply playing a game even though they're pretty much 'safe'. Are you overclocking the CPU? by that I mean a fixed clock and/or core voltage.

 
Last edited:

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
That's a good mounting arrangement. Any air should settle in up at the top tank and never be able to make it back into the pump.

Yeah...those average temps seem a bit high to me for simply playing a game even though they're pretty much 'safe'. Are you overclocking the CPU? by that I mean a fixed clock and/or core voltage.

Screenshots from my BIOS with relevant settings on screen.

EDC limit down to 110 is a new hint I picked from another suggestion somewhere else who mentioned that 134a was a bit too hungry.

Also Negative offset per core via Curve Optimizer is on where the fastest cores according to Ryzen Master are taken down by 10 and the rest by 30 to see if it helps with heat.



 
Screenshots from my BIOS with relevant settings on screen.

EDC limit down to 110 is a new hint I picked from another suggestion somewhere else who mentioned that 134a was a bit too hungry.

Also Negative offset per core via Curve Optimizer is on where the fastest cores according to Ryzen Master are taken down by 10 and the rest by 30 to see if it helps with heat.



I've found EDC setting is pretty important for determining how the CPU performs with curve optimizer. If I set it at the stock setting for my 5800X...145A...the CPU boosts agressively when cool but doesn't seem to hold clocks as well in long workloads. That shows in Cinebench scores where I can get up to 50 points higher with EDC set to 120A. I think the higher EDC setting lets it boost higher early but that also gets the cores hot so it pulls back clocks earlier and further and it never recovers.

You can also use PPT and TDC to adjust how aggressively it's boosting to control temps. But the easiest way is to simply set a platform thermal limit to the point you feel comfortable. But that's pretty much a last resort, if you're absolutely confident the Galahad is working as it should be. But always keep in mind that 90C is expected and by design. I'm just not sure a game should be working the CPU so hard it should get there.
 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
I've found EDC setting is pretty important for determining how the CPU performs with curve optimizer. If I set it at the stock setting for my 5800X...145A...the CPU boosts agressively when cool but doesn't seem to hold clocks as well in long workloads. That shows in Cinebench scores where I can get up to 50 points higher with EDC set to 120A. I think the higher EDC setting lets it boost higher early but that also gets the cores hot so it pulls back clocks earlier and further and it never recovers.

You can also use PPT and TDC to adjust how aggressively it's boosting to control temps. But the easiest way is to simply set a platform thermal limit to the point you feel comfortable. But that's pretty much a last resort, if you're absolutely confident the Galahad is working as it should be. But always keep in mind that 90C is expected and by design. I'm just not sure a game should be working the CPU so hard it should get there.

Actually I am wondering if this motherboard is working as it should be.
 
illustration below for the orientation of fans
if those are two bottom mounted exhaust fans you should turn them to intake.

you have created a negative air pressure with this scenario that will cause overall dust buildup by pulling in through every available crack or crevice in the case.
and will also keep any cooler air from properly circulating over the entirety of the system by whisking it away as soon as it passes through the radiator.

having those two bottom as intake will also provide more cool air for the GPU.
they hit well above 90
another change you can try that will sometimes lower CPU temp with a radiator is using a 6x push/pull setup of fans.

an old EVGA CLC 280 i used had issue keeping my 5.0GHz 8700K under 65°C while gaming.
adding a set of Cryorig QF140 Performance fans in the pull position on the inside of the radiator helped a lot with keeping them below 60°C.

but i would agree with other members that there is either a fault with the cooler itself, with it's installation, with the BIOS settings, or with some system control software you may be running.
you are hitting way higher temperatures than this setup should allow when setup correctly.
Screenshots from my BIOS
and 1.4v is very high for 3.7GHz:
bios3.jpg
 
Last edited:
I started trying stuff after seeing stock values do not help.
I see now, it wasn't clear at first.

I'm not sure how 'good' the Lian Li coolers are but I do know some others have had bad quality history. A common problem is the liquid harbors growth of anaerobic slime mold that can clog the micro-channels in the water block. That happens if the mfr. fails to include a biocide in the liquid coolant mix. Usually, though, you see that as temps that start off OK but climb steadily even while only in idle mode. But I suppose a less severe affliction might show similar to your experience.
 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
I see now, it wasn't clear at first.

I'm not sure how 'good' the Lian Li coolers are but I do know some others have had bad quality history. A common problem is the liquid harbors growth of anaerobic slime mold that can clog the micro-channels in the water block. That happens if the mfr. fails to include a biocide in the liquid coolant mix. Usually, though, you see that as temps that start off OK but climb steadily even while only in idle mode. But I suppose a less severe affliction might show similar to your experience.


Well, I watched lots of Gamers Nexus videos before I went ahead with Lian Li Galahad 360. What you say might be true - the bad quality history, after all GN would not be able to detect such an aftersales status.

I see I have to make sure that Lian Li is working as it is supposed to from all I have gathered here so far.
 
...
I see I have to make sure that Lian Li is working as it is supposed to from all I have gathered here so far.
The best way to do that is to swap it out with another CPU cooler I suppose. Hopefully you have access to one.

An interesting video...someone troubleshooting a cooling problem and turns up an MSI AIO with this very problem. MSI's AIO's have earned a bad reputation, maybe not for this but it's quite relevant.

View: https://www.youtube.com/watch?v=9U_5aGf0G9Q
 
Last edited:

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
The best way to do that is to swap it out with another CPU cooler I suppose. Hopefully you have access to one.

An interesting video...someone troubleshooting a cooling problem and turns up an MSI AIO with this very problem. MSI's AIO's have earned a bad reputation, maybe not for this but it's quite relevant.

View: https://www.youtube.com/watch?v=9U_5aGf0G9Q

I am riddled with problems now - I had a Scythe Ninja 5 lying around, will be swapping coolers soon. However since Scythe Ninja 5 is not famous for its thermal performance, I am not sure I'll be able to draw a comparison sustainable enough to isolate the culprit. I got no other choices left though. Thanks.
 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
Removed Lian Li completely, installed back Scythe Ninja 5 air cooler of which intake/exhaust pair of fans could ran only at 800 rpm max and below is the screenshot I got while I was playing Warzone with BIOS defaults.

 

Pretended

Distinguished
Sep 15, 2014
36
0
18,530
Might be a conclusion to this thread as I provided documentation to Lian Li and they have sent an email back today saying that they will be sending me a pump and a radiator for replacement.

Thank you all who tried to help.